Re: Optimal Flexible Architecture (OFA)

From: Howard J. Rogers <hjr_at_dizwell.com>
Date: Sun, 16 Nov 2003 10:50:08 +1100
Message-ID: <3fb6bbbb$0$20418$afc38c87@news.optusnet.com.au>

"Daniel Morgan" <damorgan_at_x.washington.edu> wrote in message news:1068939309.645607_at_yasure...

> You are focusing on one aspect of OFA that, I know, is a big issue
> in this usenet group due to the promotors of mythology. But OFA is
> also, and I'd argue far more importantly, the way home directories
> are laid out.

I wasn't focusing on anything in particular. As I said, I/O contention is one aspect (dealt with by RAID) -and it's not all mythology, sinc SYSTEM and UNDO and TEMP should not be sharing the same device, except that in RAID you have no say in the matter, and probably don't need a say anyway. But the other aspects I also dealt with. Including the one of multiple homes... which ideally shouldn't arise, as I said. But if it does, then I fail to see why /product/oracle/8.1.7 and /product/oracle/9.2.0 is any better than /8.1.7 and /9.2.0.

The one saving grace of OFA is that it would be the same on multiple installations, and you would know where to expect things to be. I can't put a price on that, so the discussion may be moot.

But OFA was only ever commonsense writ large, and provided *someone's* commonsense applies, I see no reason to slavishly adopt OFA these days. OFA *principles*, yes. OFA itself... not necessarily.

Regards
HJR Received on Sat Nov 15 2003 - 17:50:08 CST